This is an interesting idea, but I think distinct from our light cone generator.  Eric is right that they are not accounting for evolution of the structures.  They are only using one data cube to make the image.  In theory, I think it could be possible to use different data outputs for each of the stripes so as to actually sample data from the corresponding redshift.  However, this may lead to duplicate objects since many of the structures are moving and may not be in the same place in a different data output.

I can really only see this being useful for making the bowtie-like diagrams as in the final figure of the paper, but they are pretty cool.  If someone wants to try to integrate their code, it certainly wouldn't be a bad thing.


On Wed, Mar 17, 2010 at 7:33 AM, Eric Hallman <> wrote:
This is very interesting.  On a first read through though, I'm still not entirely clear what they're doing.  First of all, I am not sure how they account for the redshift evolution of portions of the universe further away.  Certainly the main goals (not using any part of the volume more than once, preserving the periodicity of the box) are important, and can be difficult to solve. I need to read through this carefully to see how this could be used for our light cones.

On Mar 16, 2010, at 10:30 PM, Matthew Turk wrote:

Hi guys,

There was a paper on the arxiv today,

about remapping cubical, periodic volumes into non-cubical volumes --
suitable for things like light cones and whatnot.  The group provides
code as well as further description here:

I wonder if this could be used as an enhancement in the light cone
extension?  Britton, Eric, what are your thoughts?

Yt-dev mailing list

Eric Hallman
Google Voice: (774) 469-0278

Yt-dev mailing list